Molecular biotechnology has revolutionized the production of protein compounds of pharmacological importance. The advent of recombinant DNA technology permitted for the first time the production of proteins on a large scale in a recombinant host cell rather than by the laborious and expensive isolation of the protein from cells or tissues which may contain minute quantities of that protein. The production of proteins, including human proteins, on a large scale in a host requires the ability to express the protein of interest in a host cell, e.g., a heterologous host cell. This process typically involves isolation or cloning of the gene encoding the protein of interest followed by transfer of the coding region (open reading frame) into an expression vector which contains elements (e.g., promoters) which direct the expression of the desired protein in the host cell. The most commonly used means of transferring or subcloning a coding region into an expression vector involves the in vitro use of restriction endonucleases and DNA ligases. Restriction endonucleases are enzymes which generally recognize and cleave a specific DNA sequence in a double-strand DNA molecule. Restriction enzymes are used to excise a DNA fragment which includes a coding region of interest from the cloning vector and the excised DNA fragment is then joined using DNA ligase to a suitably cleaved vector with transcription regulatory sequences in such a manner that a functional protein can be expressed when the resulting expression vector is introduced to a cell or an in vitro transcription/translation mixture.
A problem in controlling fragment orientation in fragments generated by restriction enzymes is that many of the commonly used restriction enzymes produce termini that are rotationally equivalent, and therefore, self-ligation of DNA fragments with such termini is random with regard to fragment orientation. Hartley and Gregori (Gene, 13:347 (1981)) reported a technique to control fragment orientation during ligation, which required the introduction of AvaI sites flanking either end of the cloned fragment (also see Hartley and Gregori, U.S. Pat. No. 4,403,036). Since AvaI cleavage produces distinguishable ends, self-ligation of the fragment results in a strong bias toward head-to-tail orientation. This is so because head-to-head and tail-to-tail ligation results in base mismatches. The polymerized molecules were then inserted into a vector and used to transform E. coli. 
In a similar approach, Ikeda et al. (Gene, 71:19 (1988)) produced head-to-tail tandem arrays of a DNA fragment encoding a human major histocompatibility antigen that was flanked by SfiI cleavage sites. SfiI produces single-strand DNA overhangs that are not rotationally equivalent. SfiI sites have also been used to produce copolymers of gene expression cassettes and selection markers, which can be used to transfect cells (Monaco et al., Biotechnol. Appl. Biochem., 20:157 (1994); Asselbergs et al., Anal. Biochem., 243:285 (1996)). Monaco et al. treated the copolymer with NotI to cleave the DNA at the 3′ end of the selectable marker gene. In this way, transfected DNA molecules contain only one selectable marker gene per copolymer.
Class IIS restriction enzymes can generate totally asymmetric sites and complementary cohesive ends. Kim and Szybalski (Gene, 71:1 (1988)) introduced sites for BspMI, a class IIS restriction enzyme, at either end of cloned DNA. Self-ligation of the cloned DNA provided multimers comprising repeat units in the same orientation. Similarly, Takeshita et al. (Gene, 71:9 (1988)) achieved tandem gene amplification by inserting a fragment encoding human protein C into a plasmid to introduce asymmetric cohesive ends into the fragment. In this case, sites for the class IIS enzyme, BstXI, were used. The multimer was then cloned into a cosmid vector comprising a neo gene, packaged into lambda phage particles, and amplified in E. coli. The cosmid vectors were then introduced into Chinese hamster ovary DHFR-cells, which were treated with G418 to select for cells that expressed the neo gene. Takeshita et al. also found that cells expressed human protein C, albeit at lower levels, following transfection with unpackaged tandem ligated DNA comprising copies of the cosmid vector and the human protein C gene.
A similar approach was described by Lee et al. (Genetic Analysis: Biomolecular Engineering, 13:139 (1996)), who amplified target DNA as tandem multimers by cloning the target DNA into a class IIS restriction enzyme cleavage site of a vector, excising a monomeric insert with the class IIS restriction enzyme, isolating monomeric inserts, self-ligating the inserts, and cloning the multimers into a vector. According to Lee et al., such a method is useful for polymerizing short DNA fragments for the mass production of peptides.
Another approach for forcing directional ligation is to devise synthetic linkers or adapters that are used to create asymmetric cohesive ends. For example, Taylor and Hagerman (Gene, 53:139 (1987)) modified the Hartley-Gregori approach by attaching synthetic directional adapters to a DNA fragment in order to establish control over fragment orientation during ligation. Following polymerization, the multimers were ligated to a linearized vector suitable for E. Coli transformation. Stahl et al. (Gene, 89:187 (1990)) described a similar method for polymerizing DNA fragments in a head-to-tail arrangement. Here, synthetic oligonucleotides were designed to encode an epitope-bearing peptide with 5′-protruding ends complementary to the asymmetric cleavage site of the class IIS restriction enzyme, BspMI. After polymerization, the peptide encoding fragments were inserted into the unique BspMI site cleavage site of a vector, which was used to transform E. coli. Clones were screened using the polymerase chain reaction, and then subcloned into prokaryotic expression vectors for production of the peptides in E. coli. 
Nevertheless, the ability to transfer a desired coding region to a vector with transcription regulatory sequences is often limited by the availability or suitability of restriction enzyme recognition sites. Often multiple restriction enzymes must be employed for the removal of the desired coding region and the reaction conditions used for each enzyme may differ such that it is necessary to perform the excision reactions in separate steps. In addition, it may be necessary to remove a particular enzyme used in an initial restriction enzyme reaction prior to completing remaining restriction enzyme digestions. This requires a time-consuming purification of the subcloning intermediate. It also may be necessary to inactivate restriction enzymes prior to ligation.
Methods for the directional transfer of a target DNA molecule from one vector to another in vitro or in vivo without the need to rely upon restriction enzyme digestions have been described. For example, the Creator™ DNA cloning kit (Clontech Laboratories, Inc.) uses Cre-loxP site-specific recombination to catalyze the transfer of a target gene from a donor vector to an acceptor vector, which is a plasmid containing regulatory elements of the desired host expression system (see also U.S. Pat. No. 5,851,808). Cre, a 38-kDa recombinase protein from bacteriophage P1, mediates recombination between or within DNA sequences at specific locations called loxP sites (Sauer, Biotechniques, 16:1086 (1994); Abremski et al., J. Biol. Chem., 259:1509 (1984)). These sites consist of two 13 bp inverted repeats separated by an 8 bp spacer region that provides directionality to the recombination reaction. The 8 bp spacer region in the loxP site has a defined orientation which forces the target gene to be transferred in a fixed orientation and reading frame. Donor vectors in the kit contain two loxP sites, which flank the 5′ end of a multiple cloning site (MCS) and the 5′ end of the open reading frame for the chloramphenicol resistance gene. Donor vectors also contain the ampicillin gene for propagation and selection in E. coli, and the sucrase gene from B. subtilis (SacB) for selection of correct recombinants. Acceptor vectors in the kit contain a single loxP site, followed by a bacterial promoter, which drives expression of the chloramphenicol marker after Cre-lox-mediated recombination. The gene of interest, once transferred, becomes linked to the specific expression elements for which the acceptor vector was designed. If the coding sequence for the gene of interest is in frame with the upstream loxP site in the donor vector, it is in frame with all peptides in the acceptor vector.
The Gateway™ Cloning System uses phage lambda-based site-specific recombination. The LR Reaction is a recombination reaction between an entry clone having mutant attL sites and a vector (a Destination Vector, pDEST™) having the corresponding mutant attR site, mediated by a cocktail of recombination proteins (λ recombination proteins Int, Xis, and the E. coli-encoded protein IHF), to create an expression clone. The BP Reaction is a recombination reaction between an expression clone (or an attB-flanked PCR product) and a donor vector to create an entry clone. The BP reaction permits rapid, directional cloning of PCR products synthesized with primers containing terminal 25 bp attB sites (+4 Gs). The result is an entry clone containing the PCR fragment. Similarly, DNA segments flanked by attB sites in an expression clone can be transferred to generate entry clones which can be used to move the sequence of interest to one or more destination vectors in parallel reactions to generate expression clones. The resultant 25 bp attB sites (attB1 on the left (N-terminus) and attB2 on the right (C-terminus)) created by the LR reaction are derived from the attL sites (adjacent to the gene), whereas the distal sequences are derived from the attR sites.
However, the protein encoded by Cre-loxP based expression vectors or other site-specific recombinase based vectors, e.g., the Gateway™ Cloning System, has numerous, for instance, 8 to 13, amino acid residues at the N-terminus and C-terminus of the protein, which residues are encoded by the site-specific recombination exchange sites.
Thus, what is needed is an improved method to directionally clone a nucleic acid sequence of interest.